Two central goals of health policy are to improve population health as much as possible and to distribute the improvements fairly. These goals will often conflict. Reasonable people will disagree about how to resolve these conflicts, which take the form of various unsolved rationing problems. The conflict is also illustrated by the ethical controversy that surrounds the use of cost-effectiveness analysis. Because there is no consensus on principles to resolve these disputes, a fair process is needed to assure outcomes that are perceived to be fair and reasonable. One such process, accountability for reasonableness, assures transparency, involves stakeholders in deliberating about relevant rationales, and requires that decisions be revised in light of new evidence and arguments. It has been influential in various contexts including developed countries such as Canada, the United Kingdom, New Zealand, and Sweden, and developing countries, such as Mexico.
Decisions about health-care provision are morally contentious because problems of distributive fairness are pervasive. We lack agreement on principles of distribution fine-grained enough to resolve the many disputes we have about how to balance maximization of aggregate measures of population health against the fair distribution of health benefits, yet both goals—improving population health and doing so fairly—are the primary goals of health policy. In the face of these pervasive and systematic controversies, a reasonable health policy will be one that emerges from a fair, deliberative process of the sort that holds decision-makers accountable for their reasonableness. Such a process must be based on reasons people can agree are relevant to the task, and on rationales that are fully transparent to those affected by the decisions and that allow for the revision of decisions in light of new evidence and arguments.
This process-based approach will leave some people unsatisfied: it does not tell us exactly how much a country should be spending on health care, for example, and it does not tell us exactly how what it does spend should be distributed. The answers to those questions must emerge from the fair process itself. What the process assures us of, however, is that the moral controversies about distribution—how to trade-off maximizing benefits against their equitable distribution—are thoroughly considered and that choices reflect careful consideration of the different values involved. That may be the best we can do in the pursuit of our two main goals, improving population health and distributing the improvements fairly.
Improving population health as much as we reasonably can is a requirement of justice because the departures from normal functioning that constitute ill health interfere with the satisfaction of various principles of justice. Thus ill health diminishes significantly the range of exercisable opportunities open to people, as I have argued elsewhere, or, as other philosophers have claimed, it diminishes their capabilities or their opportunity for welfare and advantage. Accordingly, different, prominent accounts of justice emphasize the importance of establishing institutions that protect or promote these goods (which is not to say that some theories deny any obligations to do so). Distributing those improvements fairly, however, is a further requirement of justice. A reasonable health policy must address and resolve the conflicts that may arise between these goals.
Since the conflict between these goals is so common, it may seem ironic that it ultimately disappears, at least if the goals are fully achieved. Health, conceived of as normal functioning, is a concept with a limit, unlike wealth or income. Consequently, the ultimate goal of maximizing population health is achieved if all people are made fully functional or healthy. But that outcome also fairly distributes the improvements, since all end up equally healthy. Ultimately, and surprisingly, the health maximizer and the health egalitarian (i.e. someone concerned about equity in health outcomes) aim for the same goal. In the real world where many are unhealthy, however, the goals rudely come apart. Health-maximizing strategies often ignore equity in health, and the pursuit of equity in health often requires foregoing some aggregate health benefits. It is in this uneven terrain that we must seek reasonable ways of balancing our two goals.
Reasonable people will, unfortunately, disagree about how to balance these objectives. In the absence of a prior, principled agreement that resolves these disputes, we require a form of procedural justice, a process accepted as fair by all. Then, outcomes of the fair process will be accepted as fair. A reasonable health policy will be what results from such a process, for the process holds decision-makers accountable for the reasonableness of their decisions. We illustrate more carefully in the next section why such a fair process is needed, but first one preliminary point needs to be addressed.
Some people might believe the problem of resolving these disagreements about priorities in health policy is artificially created by the presence of resource limits that are avoidable. They believe we should be able to do everything for everyone in need, and if we eliminated waste or arbitrary political constraints on funding, then we could meet all such needs. Some form of resource scarcity, however, is always going to be present: promoting health is not the only important good, for every society must meet other needs of its citizens. In addition, new technologies and ageing populations with more expensive health needs mean that we are always facing resource allocation problems with some novelty. Therefore, even very inefficient health delivery systems, such as the highly fragmented American system with its uniquely high unit costs and administrative costs, would face scarcity problems not due to waste were it to be dramatically reformed on the model of universal coverage systems elsewhere.
Health maximization vs health equity
To illustrate the types of conflict between health maximization and health equity that make health policy decisions difficult, consider one standard methodology for maximally improving population health; namely, cost-effectiveness analysis (CEA). The method permits us to compare the incremental health benefits of an intervention or policy with its incremental costs. The ratio of the two is the incremental cost-effectiveness ratio. Comparing the incremental cost-effectiveness ratios of two treatments for the same condition or two treatments for different conditions allows us to measure the marginal health benefit gained per unit of money spent.
Health benefits can be measured in various ways, but some way of combining the duration and the quality of changes in health states into health-adjusted life years (HALYs) is preferred so one can make comparisons across conditions with different kinds of outcomes. The most commonly used such unit in medicine is the quality-adjusted life year (QALY). Different health states are assigned a value on a scale from 0 (death) to 1 (perfect health), and this weighting function is then multiplied by the duration of live lived in that state to yield the number of QALYs gained. Thus, a year lived in perfect health is equivalent to 1 QALY, but if one is in a significantly compromised health state having a value of 0.6, then that same year is equivalent to 0.6 QALYs. The costs are similarly limited to treatment costs and do not include the indirect costs that result if people live longer and have to be assisted in other ways. If we select the most cost-effective intervention for treating a specific condition, we are getting the most health benefit per unit of money spent.
Although this method is often characterized as ‘utilitarian’ because it is a maximizing strategy, unlike utilitarianism, it ignores benefits other than the health benefit to the individual—for example the income stream generated by the extra life years the individual enjoys—and it ignores costs other than treatment costs. Arguably, some of these excluded benefits and costs are relevant to social policy, so cost-effectiveness analysis and full-blown utilitarianism will differ in some policy recommendations. Still, one ethical reason for avoiding some of these other benefits and costs is that it allows policy in the medical sector to avoid making social worth judgements about the consequences of treating some people rather than others. Arguably, this kind of respect for persons is owed to people in need of medical interventions and is part of the rationale for treating medicine as something of a ‘separate sphere’ from some other policy domains.
Despite this effort to keep CEA free from certain moral objections, it still faces other ethical criticisms. A key form of criticism is that the maximizing commitments of CEA mean it ignores important issues of distributive fairness. Consider three unsolved rationing problems in which CEA adopts an extreme position that puts it at odds with widely held concerns about equity.
◆ The priority problem. One widely held view is that we should give some priority in our thinking about whom to treat to those who are worse off. If, however, we give complete priority to those who are worse off, following a ‘maximin’ rule, then we may have to forego very significant benefits to others who have legitimate claims to them while we secure only minor benefits for the worst off. Maximin seems implausible. CEA, however, tells us to give no priority to those who are worse off. Rather, we are to distribute the chips—the QALYs—where they fall, aiming only to get the most chips. In effect, CEA denies or ignores the widely held view that it may be more morally important to give a QALY to someone who is worse off than another. But if both extreme views are implausible—no priority and maximal priority—then how much priority should we give those who are worse off? This is the priority problem, and reasonable people disagree about it.
◆ The aggregation problem. CEA tells us that all aggregations of benefit are permissible: we seek the maximum aggregate benefit regardless of whether it arises from significant benefits to a few or from very modest benefits to many. In contrast, most people think some trivial additional benefits should not count in deciding whom to save. At the other extreme, some people will rule out all forms of aggregation. Most people, however, believe that it is preferable to save more lives or life years or QALYS than fewer, even though they reject allowing every conceivable aggregation. Again, extreme positions seem implausible but reasonable people will disagree about which aggregations to allow in the middle range.
◆ The best outcomes/fair chances problem. As in the other problems, CEA takes an extreme view, always favouring best outcomes (greatest health benefit). But always favouring a group that would achieve best outcomes, say those who will gain the most life-years or QALYs from a transplant, means that others may have no chance at all at a significant benefit, say living more years but not as many as those in the ‘best outcome’ group. They would complain they ought to have a fair chance of some benefit and not be asked to sacrifice any chance of significant benefits just because others are lucky enough to have better outcomes. An extreme alternative to always favouring ‘best outcomes’ is to give everyone equal chances of a benefit. Many would think it implausible to give people equal chances of a benefit when that would mean foregoing much greater benefits for others. To find a middle ground, some theorists have therefore argued for a weighted or proportional lottery or chance. Reasonable people will disagree about how to set those weights or proportions.
Notice that standard policy choices about health interventions often face one or another of these problems. For example, in decisions to scale up antiretroviral treatments in high-prevalence, low-resource countries, it is often claimed that locating treatment sites in urban hospitals would lead more quickly to higher treatment rates. The decision means, however, that other populations fail to have any chance at significant benefit because ‘best outcomes’ favours another group. The problem of orphan drugs has the same form. Similarly, since CEA favours best outcomes, then wherever more QALYs are gained by treating people who are young than those who are old, the old lose any chance at a significant benefit while the young are favoured. (Some may want to give priority to the young for other reasons of fairness, such as Williams’s ‘fair innings argument’, but always favouring best outcomes is not a fairness consideration.)
These distributive problems, and thus the conflict between health maximization and health equity, arise in two important and perhaps surprising contexts, efforts to reduce health inequities and human rights efforts to promote health. The distributive problems just reviewed are all presented with a morally unproblematic baseline. Some people are sicker than others, but we do not ask why. We only consider how much priority to give to them compared to others. Suppose, however, that some group is unjustly worse off with regard to health than another. Perhaps it is a traditionally excluded group, such as the indigenous population in some Latin American countries that has high maternal mortality rates, or African-Americans in the United States of America who have a higher prevalence of uncontrolled high blood pressure, or some ethnic minorities, like the Maori in New Zealand or Aboriginals in Australia, or the Roma in some European countries, all of whom have higher infant mortality rates than the dominant population, or young women in sub-Saharan Africa who have higher prevalence of HIV/AIDS than men. Because of the injustice of the baseline, we may have some good reason to want to give more priority to helping those who are worse off than we ordinarily might have, but how much more? Giving maximum priority (‘maximin’) means we forego all other benefits to others, whose illness gives them legitimate claims on assistance even it if is not the result of some broader social injustice. We may shift our priorities, but reasonable people will again disagree how much.
One promising movement in international health, the broad effort to improve population health and its distribution through the international legal framework of human rights, also encounters these controversial distributive problems. We might think this implausible because there is already agreement on general principles, the specific rights that are included in this framework. Nevertheless, we can often only improve realization of a specific right for some claimants and not others. For example, we may reach some young girls but not others in a programme to improve literacy among young women, a key determinant of maternal and child health. We can often improve population health in various ways that improve realization of some rights and not others, say by improving training of birth attendants or by changing the legal system to prevent young girls from being married and becoming pregnant too early. The choices among these alternatives raise the same priority, aggregation, and best outcomes/fair chances problems we already encountered, and the language of rights does not eliminate them. This means that different people affected by human rights efforts to improve population health will disagree about the fairness of one reform programme as opposed to another.
Accountability for reasonableness
When we lack prior agreement on principles fine-grained enough to resolve moral disagreements about fair distributions, we must rely on a form of procedural justice. For complex health policy choices, we need a process that is fair to all stakeholders, that is transparent to the broader public affected by these decisions, that permits appropriate deliberation about the reasons that are relevant to choices, and that allows decisions to be revisited in light of new evidence and arguments. One version of such a process is called ‘accountability for reasonableness’, and it requires compliance with the following conditions:
◆ Publicity. The rationales for all decisions must be publicly accessible so that people affected by them understand why choices that fundamentally affect their well-being were made.
◆ Relevant reasons. The rationales for decisions must be based on reasons that fair-minded people agree are relevant to making decisions of this sort. Fair-minded people are people interested in being able to justify their decisions to each other. Often, a mechanism for vetting such reasons is to involve a broad group of stakeholders in the decision, though the exact form for such involvement may vary considerably with the institutional level at which decisions are made. Stakeholder involvement should not be understood as ‘democratic’; rather, it improves the deliberation and helps to assure a careful assessment of all relevant reasons, as well as adds transparency to the deliberation. (Stakeholder involvement may not be feasible in some private organizations, and then the publicity condition is the only assurance that broader critical discussion can take place.)
◆ Revisability and appeals. Decisions must be evidence and argument-based, and this means a fair process must allow for revisiting decisions as new considerations become available. Further, there are often specific groups or individuals who do not quite fit the rules or reasoning underlying some choices and who are unduly burdened by them. Such people need a fair hearing of their appeal against decisions that deny them fundamental goods.
◆ Enforcement. There needs to be assurance, often in the form of regulation, makes sure the procedures are ones that comply with the first three substantive conditions.
If these conditions are met, decision-makers are held accountable for the reasonableness of their decisions. Indeed, a kind of case law emerges that reveals the commitments underlying the decisions. In this way, the reasons underlying decisions are made explicit, even if all the grounds for them cannot all be articulated prior to considering specific cases, and they gradually accumulate as an explicit framework for priority setting. Since the public reporting of full rationales shows that deliberation about relevant reasons is at the core of the process, the public is assured that moral reasoning plays a role in what have historically been ‘behind the scenes’ decisions made by bureaucrats unaccountable for their reasons. The public reporting also means that morally controversial choices are exposed to broader public review and criticism. Ultimately, the fair process not only institutionalizes the importance of moral reasoning about these problems but contributes to a social learning process that can improve democratic regulation of the institutions over time.
Accountability for reasonableness is intended as a fair procedure for resolving disagreements about some problems of resource allocation problems, but it is constrained by some substantive moral commitments. For example, human rights considerations would restrict decisions about priorities to ones that are, among other things, nondiscriminatory. It is also assumed that there is a shared goal of meeting health needs fairly under resource constraints while promoting population health. More specifically, the shared goal might be characterized as a goal of protecting opportunity or capabilities by protecting health—and then more specific reasons for constraints must be the subject of deliberation in the fair process. Thus reasonable people may disagree about exactly what role to assign responsibility for healthy behaviours, especially given the difficulty of assigning responsibility, and this matter will have to be resolved along with others. A further issue that affects the legitimacy of outcomes of the fair process is specification of who should have the moral authority to make decisions. Much depends on the level within a health system at which decisions are made and the nature of the organization and funding of the system: it will be far more appropriate in publicly administered systems to be more inclusive about stakeholders involved in or consulted on decisions than it can be in systems that have private, indeed for-profit, components. In the latter, the publicity condition is an effort to include the broader public in the political oversight of the decisions as a whole.
A fair process such as this has broad appeal and broad application to many kinds of health care decisions at different institutional levels: international, national, and subnational. Internationally, the World Health Organization recently endorsed the value of such a process in its guidelines for assuring equity in patient selection for the scaling-up of antiretroviral treatments in its ‘3 by 5’ programme. It also uses the features of that process as a framework for assessing decision-making in Tanzania about its AIDS treatment policy. At a national level, the Ministry of Health in Mexico recently embraced the features of accountability for reasonableness in constructing a fair, deliberative process for making decisions about incremental expansion of its catastrophic insurance plan. In the United States of America, an Institute of Medicine committee reporting on the use of CEA in regulatory settings insisted that the method, while valuable, cannot (by design and use) capture all value issues, so it should be one input among others into a fair deliberative process of the sort that meets the conditions of accountability for reasonableness. In the United Kingdom, Sir Michael Rawlins, director of the National Institute for Health and Clinical Excellence (NICE), affirms that the ideas underlying accountability for reasonableness should inform NICE’s technology assessment recommendations. The Citizen’s Council NICE established is one version of an appeal to a broader group of stakeholders and the commitment to full publicity for the grounds of decisions is similarly a positive feature of NICE’s method of operation.
The process can be used to determine a reasonable health-care policy at subnational levels as well. Health authorities working at the regional and local levels face priority-setting decisions that encounter the same distributive problems noted earlier. Similarly, hospitals make priority and limit-setting decisions that inevitably encounter these distributive problems, and hospitals and their subdivisions could all comply with institutionally-appropriate versions of the conditions for accountability for reasonableness. Even general practitioners, acting as fundholders or commissioners of health services, would improve the legitimacy of their decisions were they to hold themselves accountable for their reasonableness.
Brock D (2003). Ethical issues in the use of cost effectiveness analysis for the prioritization of health care resources. In: Khusfh G, Englehardt T (eds.) Bioethics: a philosophical overview. Kluwer, Dordrecht, pp. 203–19. [Explores a range of criticisms of cost-effectiveness analysis.]Find this resource:
Daniels N (2008). Just health: meeting health needs fairly. Cambridge University Press, New York. [Provides a comprehensive and integrated account of justice and health.]Find this resource:
Daniels N, Sabin JE (2002). Setting limits fairly: can we learn to share medical resources? Oxford University Press, New York. [Describes the theory underlying accountability for reasonableness and explains the need for it.]Find this resource:
Gold MR, et al. (1996). Cost-effectiveness in health and medicine. Oxford University Press, New York. [This Public Health Service report articulates a standard form of cost-effectiveness analysis.]Find this resource:
Ham C, Robert G (1999). Reasonable rationing: international experience of priority setting in health care. Open University Press, Maidenhead. [Evaluates the degree to which rationing conforms to conditions of reasonableness in five countries.]Find this resource:
Kamm F (1993). Morality, mortality: death and whom to save from it, vol. 1. Oxford University Press, Oxford. [Uses rigorous philosophical tools to explore problems of priority setting in health care.]Find this resource:
Miller W, Robinson LA, Lawrence RS (eds) (2006). Valuing health for regulatory cost effectiveness analysis. National Academies Press, Washington, DC. [This Institute of Medicine report supports the view that cost-effectiveness analysis should be seen as an input into a deliberative process.]Find this resource:
World Health Organization (2006). Equity and fair process in scaling up antiretroviral treatments: potentials and challenges in the United Republic of Tanzania. WHO, Geneva. [Evaluates the Tanzanian policy for scaling up antiretroviral treatments for AIDS in light of accountability for reasonableness.]Find this resource: